IEEE SENSORS JOURNAL, VOL. 4, NO. 5, OCTOBER 2004
691
Tactile Sensor Based on Piezoelectric Resonance
G. Murali Krishna and K. Rajanna
Abstract—We discuss here the realization of tactile sensors
based on the principle of change in piezoelectric resonance frequency with the applied pressure. An array of electrodes has been
adopted on either side of the PZT material to have independent
resonators. The common areas sandwiched between the electrodes
and excitable at resonance frequency of the PZT material are used
to form the sensitive area of the tactile sensor. The electrodes were
deposited using sputtering technique. Tactile sensors with 3 3,
7 7, and 15 15 array of electrodes are developed with different
electrode dimensions and separation between the electrodes. The
tactile sensor has been interfaced to computer for the convenience
of automatic scanning and making it more user interactive. The
tactile sensors developed with different spatial resolution were
tested for different shaped objects placed in contact with the
sensor. The 3 3 matrix tactile sensor showed relatively poor
spatial resolution, whereas the 15 15- matrix tactile sensor
showed improved spatial resolution. The sensor with 7 7 matrix
elements was tested for its sensitivity to different extents of applied
force/pressure. The output response study carried out on the
sensors indicated that these sensors can provide information not
only about the extent of force/pressure applied on the object, but
also the contour of the object which is in contact with the sensor.
Index Terms—Piezoelectric resonance and robotics, tactile
sensor.
pacitive [4], magnetoresistive [7], fiber-optic [7], piezoelectric
[1], [11], and surface acoustic wave (SAW) [15] techniques. In
our present work, we have used the principle of change in the
piezoelectric resonance frequency of poled ceramic lead zirconium titanate (PZT) with the applied stress. We have adopted an
array of electrodes to have independent resonators. The tactile
sensor developed is interfaced to computer and automated to test
its output behavior. The response of the tactile sensor has been
tested for different shapes of objects placed in contact with the
sensor. The sensor is also tested for different extents of pressure
applied over the object placed in contact with the sensor.
II. EXPERIMENTAL
A. Principle
When an applied frequency to a piezoelectric material is
equal to its mechanical resonance frequency, then the material
is said to be in piezoelectric resonance. Some oscillator circuits,
when suitably connected with piezoelectric material, will give
output frequency at the mechanical resonance frequency of the
material. The resonance frequency of the resonating material is
given by the following: [2]
I. INTRODUCTION
T
OUCH and tactile sensors are devices which measure the
parameters of contact between the sensor and an object.
Touch sensing is basically the process of detection and measurement of a contact force at a defined point. A touch sensor
can also be restricted to binary information, namely touch and
no touch, whereas tactile sensing is the process of detection and
measurement of the spatial distribution of forces perpendicular
to a predetermined sensory area, and the subsequent interpretation of the spatial information. A tactile sensing array can
be considered to be a coordinated group of touch sensors [5].
Robotic and industrial automation are the potential application
areas, which have generated considerable interest for the development of these sensors. However, there are several other possible application areas for tactile sensors, namely agriculture,
food processing, bio-medical [16], [10], entertainment, and future domestic and service industries. Different types of principles have been explored in realizing the tactile sensors. They
are based on mechanical [12], [6], [14], resistive [17], [8], ca-
(1)
where
thickness of the piezoelectric material;
natural mechanical resonance frequency of first order;
stiffness constant of the material;
density of the material.
The area between the electrodes that is oscillating at the mechanical resonance frequency of the piezoelectric material is the
active area. For a given material, if the stiffness constant and
density of the material are kept constant, then the frequency becomes a sensitive parameter of thickness. When force/pressure
is applied on the active area, strain on the active area occurs and
the resonance frequency changes [3]. The change in resonance
is given by [13]
frequency
(2)
where
Manuscript received October 26, 2002; revised January 9, 2004. The associate
editor coordinating the review of this paper and approving it for publication was
Dr. Babak Ziaie.
G. M. Krishna is with the Department of Instrumentation, Indian Institute of Science, Bangalore-560 012, India, and also with the Institute of
Physics, University of Basel, Basel CH-4056, Switzerland (e-mail: murali.ghatkesar@unibas.ch).
K. Rajanna is with the Department of Instrumentation, Indian Institute of Science, Bangalore-560 012, India (e-mail: kraj@isu.iisc.ernet.in).
Digital Object Identifier 10.1109/JSEN.2004.833505
resonance frequency after the application of pressure;
change in the resonance frequency;
constant;
applied force.
and density of a given material are
If stiffness constant
constant, then from (1)
1530-437X/04$20.00 © 2004 IEEE
(3)
692
IEEE SENSORS JOURNAL, VOL. 4, NO. 5, OCTOBER 2004
Fig. 1. Block diagram of the experimental arrangement for testing the output response of the tactile sensor.
Hence
(4)
is thickness deformation. The negative sign implies
where
that, when force/pressure is applied on the resonating area,
thickness decreases, and, hence, the frequency increases. When
is far less than , then [13]
(5)
The change in frequency is directly proportional to the applied
force/pressure. Using this basic principle, an array type tactile
sensor is realized in the present work.
B. Sensor Array Design and Preparation
We have used the commercially available circularly shaped
poled PZT material for realizing the tactile sensor. The PZT
material used is of thickness 1 mm and diameter 20 mm with
MHz. Initially, the linear variation
a resonance frequency
of change in resonance frequency with applied pressure/force
for the chosen PZT samples was ascertained. Further, rectangular-shaped arrays of electrodes are deposited on either side of
the PZT material such that the array of electrodes on one side
are perpendicular to the array of electrodes on the other side. DC
magnetron sputtering technique has been used for depositing
the array of electrodes. Sputtering technique is chosen, as it is
known to provide good adhesion of the film on to the substrate.
In order to obtain the required array pattern of electrodes, precision mechanical masks were employed during deposition. These
masks were designed using AutoCAD and prepared by adopting
the chemical milling technique. Double enameled copper wires
were attached to the electrodes using temperature-controlled
soldering station. The common areas of PZT material sandwiched between the intersecting regions of the electrodes form
a matrix of independently excitable resonators at the resonance
frequency of the material (Fig. 1). By choosing one electrode
from the top side and one from the bottom, the common area
between them is excitable at resonance. This procedure can be
repeated for all the matrices of common areas between the electrodes. Each time, the matrix location and its corresponding resonance frequency need to be noted. This becomes the reference
data of the sensor before detecting any object. In the present
work, the tactile sensors with 3 3, 7 7, and 15 15 array
of electrodes have been prepared. In each case, the width of
array electrodes and the separation between them were suitably
chosen [9].
C. Testing of the Tactile Sensor
In order to study the output response of the tactile sensor,
the object to be studied was placed in contact with the sensor
and subjected to different extent of pressure/force using an arrangement [9] consisting of a micrometer with digital display
and capable of measuring the displacement/deformation up to
micron accuracy. Small pieces of silicon rubber covering
the electrodes region were placed on either side of the sensor.
These silicon/rubber pieces ensure smooth transfer of the applied force and avoid the possible shorting, especially in the
case of conducting objects between the electrode arrays. In the
present study, different sized and shaped objects were used.
Keeping the object in contact with the sensor, the entire matrix of elements are excited at resonance one after the other,
and their positions with corresponding frequencies are recorded.
The matrix of elements subjected to pressure show an increase
in frequency proportional to the extent of the pressure applied.
The difference in frequency (before keeping the object and after
keeping the object) is calculated. It is important to note that,
wherever the object is in contact with the sensor, the corresponding elements in the matrix only show an increase in the
frequency. In order to obtain a contour of the object that is in
contact with the sensitive area of the tactile sensor, a three-dimensional plot was made between the position of the matrix elements in the - axes plane and the change in frequency on
the axis. The process of doing the above operation manually
by selecting the individual matrix elements is time consuming
and tedious. Therefore, in order to make the whole process automatic and more user interactive, necessary electronic circuits
KRISHNA AND RAJANNA: TACTILE SENSOR BASED ON PIEZOELECTRIC RESONANCE
693
Fig. 2. Schematic diagram of the electronic circuit used for interfacing the tactile sensor to the computer.
and software programs were designed and implemented to interface the sensor with the computer [9]. A block diagram of
the complete experimental arrangement along with the tactile
sensor is shown in Fig. 1. It consists of analog multiplexers, a
series-resonant oscillator circuit, a PIC 16F84 micro-controller,
and a level converter for RS 232 interface to the computer. The
program written in the microcontroller initially selects a matrix element, and the selected matrix element gets connected to
the oscillator circuit. The frequency counter program will measure the oscillator output frequency and it sends the measured
frequency to computer through the RS 232 level converter. This
procedure gets repeated until the complete scanning of the entire
matrix of elements. Hence, the entire operation of scanning the
matrix of active elements, measuring the frequency and sending
the measured frequency to the computer, is controlled by the microcontroller. The software written for the measurement of the
frequency is capable of measuring the frequency with an accuHz at 1 MHz. All the tactile sensors made have been
racy of
tested for their output response using the above experimental arrangement and procedure.
The schematic diagram of the electronic circuitry used for interfacing the tactile sensor to the computer is shown in Fig. 2.
and
are 16-channel CMOS analog multiplexers
The ICs
which receive the inputs from the top and bottom array of sensor
electrodes. The outputs of these two multiplexers are connected
to the oscillator circuit (with IC ). The oscillator circuit resMHz) of the poled PZT
onates at the resonance frequency (
material used. The resonance frequency is measured at a gate
) which, in turn,
pulse of 1 s using an AND gate (with IC
receives the input from microcontroller (with IC
) for controlling the gate pulse duration of 1 s. The microcontroller (with
IC ) is a flash 8-bit CMOS controller (Microchip Company,
USA). The frequency measured by the microcontroller is fed to
Fig. 3. Flow chart of the entire software implemented.
RS 232 interface (with IC
). Fig. 3 shows the flow chart of
the necessary software implemented.
In order to have a calibration of the sensor output with reference force, a universal testing machine (INSTRON, U.K) along
with a load cell has been used. The output of the sensor resonance frequency with applied force was recorded both in ascending and descending order using a dual-beam storage oscilloscope (Hewlett Packard, Model no. 54600, 100 MHz). All the
experiments, including the output response study and calibration of the sensor output with reference force, have been carried
out at room temperature.
694
IEEE SENSORS JOURNAL, VOL. 4, NO. 5, OCTOBER 2004
Fig. 4. Response of 3
Fig. 5.
Response of 3
2 3 array tactile sensor for the object covering the entire active region.
2 3 array tactile sensor for the object located at one corner side of the active area.
III. DISCUSSION
The tactile sensors with different number of arrays of electrodes namely 3 3 arrays, 7 7 arrays, and 15 15 arrays
were studied for their output response by using the aluminum
plate (both square and circular shaped) as an object placed over
the sensor. In all the figures (Figs. 4–11) which show the tactile sensor response, dashed lines on the left side of the figures
signifies the number of electrodes and their location on the PZT
material (circular in shape) constituting the sensitive area, with
hashed regions indicating the location of the object over this
sensitive area and the right side of the figures show the corresponding response. Fig. 4 shows the response for the object covering the entire sensitive area of the 3 3-array tactile sensor.
The response corresponding to regions between the active elements is linearly interpolated from the data obtained from the
active elements of the sensor. As can be seen in Fig. 4, the response is higher at the center compared to the regions toward the
edges. This is thought to be due to several possible reasons, like
nonuniformity of applied pressure, nonuniform poling of the
PZT material, or it could be due to nonuniformity in thickness of
the PZT ceramic material used. The ambiguity in nonuniformity
in applied pressure can be resolved by testing the sensor with
higher spatial resolution (more number of matrix elements). The
response of the 3 3 array sensor for the object located at different regions of the active area are shown in Figs. 5 and 6. As
can be seen, it is clearly evident that the response of the sensor
is largely confined to the area covered by the object.
The response of the 7 7 array tactile sensor in case of the
square shaped object located toward one corner side of the active
area is shown in Fig. 7. The response of the sensor for the same
object located at the same position but with the increased extent
of pressure is shown in Fig. 8. The change in frequency is more
for the higher extent of applied pressure indicating the pressure
sensitivity of the sensor. Also, it is clearly evident that, in comparison with the 3 3 array sensor, the spatial resolution of 7 7
array tactile sensor has improved significantly as expected.
The response of the 7 7 array tactile sensor for a circular
object located at the center of the active area is shown in Fig. 9.
It can be observed that the response of the sensor is not only
confined to the actual area of the object which is in contact with
the sensor, but also to the closer adjacent regions beyond the
edge boundaries of the object. This is attributed to the fact that,
when pressure is applied over the object, it is transferred to the
active area through the silicon rubber. Since the silicon rubber is
elastic, it expands under the application of pressure and covers
the adjacent additional regions of the active area of the sensor.
KRISHNA AND RAJANNA: TACTILE SENSOR BASED ON PIEZOELECTRIC RESONANCE
Fig. 6. Response of 3
Fig. 7.
Response of 7
695
2 3 array tactile sensor for the object located at another corner side of the active area.
2 7 array tactile sensor for a square shaped object located toward one corner of the active area.
2
Fig. 8. Response of the 7 7 array tactile sensor for a square shaped object located toward one corner side of the active area (same as in the case of Fig. 7) and
with increased extent of pressure/force.
The response of the tactile sensor with 15 15 array of electrodes is shown in Fig. 10 for a circular object placed at the
center of the active area. Also, the response of the same sensor
for an object having a hole at the center is shown in Fig. 11. It is
clearly noticeable that, compared to the 7 7-array sensor, the
spatial resolution has further improved. It is also evident from
Figs. 10 and 11 that the sensor can provide information about
contour of the object.
Fig. 12 shows the variation of change in resonance frequency
of the sensor with applied force (in ascending and descending
696
IEEE SENSORS JOURNAL, VOL. 4, NO. 5, OCTOBER 2004
Fig. 9. Response of 7
Fig. 10.
2 7 array tactile sensor for a circular object located at the center of the active area.
Response of the 15
Fig. 11.
2 15 array tactile sensor for a circular object placed at the center of the active area.
Response of the 15
2 15 array tactile sensor for an object having a hole at the center.
order) for three consecutive cycles. It can be seen that, in all the
three cases, the output is repeatable and the calculated hysteresis
is found to be less than 1% full-scale output. It is observed that
the initial offset frequency of the sensor reverts to its original
value after each calibration cycle. Considering 0.4-N force resolution of universal testing machines used and 0.05% resolution in oscilloscope and related signal conditioner circuitry employed in our experiments, the calculated uncertainty of mea-
surement is below 0.1%. Based on the drop ball method experiment, the dynamic response of the sensor is found to be 0.05 s.
IV. CONCLUSION
We have successfully developed the tactile sensor based on
the principle of piezoelectric resonance. An array of electrodes
geometry has been employed in the realization of the sensor.
KRISHNA AND RAJANNA: TACTILE SENSOR BASED ON PIEZOELECTRIC RESONANCE
Fig. 12. Variation of change in resonance frequency with applied force for
three consecutive cycles.
The necessary electronics (both hardware and software) have
been developed and the entire assembly is interfaced suitably
with the computer. The output response study carried out on the
sensor clearly demonstrate that the sensor can provide information not only about the extent of force/pressure applied on the
object, but also the contour of the object which is in contact
with the sensor. This scope exists for improving the sensor behavior in terms of its spatial resolution by increasing the number
of array electrodes using lithographic techniques. The presently
developed tactile sensor has potential application possibilities
in the areas of robotics and biomedical engineering.
ACKNOWLEDGMENT
The authors would like to thank the Chairman, Department of
Instrumentation, IISc, for his kind support and encouragement
in carrying out this work. They would also like to thank Dr. M.
M. Nayak, LPSC and ISRO, for his kind help in many ways
and for the fruitful discussions related to this work, and Dr. L.
Shivalingappa, for his help in the early stages of this work.
REFERENCES
[1] M. F. Barsky, D. K. Lindner, and R. O. Claus, “Robot Gripper control
system using PVDF piezoelectric sensors,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. UFFC-36, pp. 129–133, Jan. 1989.
[2] E. Benes, M. Groschl, W. Burger, and M. Schmid, “Sensors based on
piezoelectric resonators,” Sens. Actuators A, vol. 48, pp. 1–21, 1995.
[3] C. Gehin, C. Barthod, and Y. Teisseyre, “Design and characterization of
new force resonant sensor,” Sens. Actuators A, vol. 84, pp. 65–69, 2000.
[4] H. Morimura, S. Shigematsu, and K. Machinda, “A novel sensor cell architecture and sensing circuit scheme for capacitive fingerprint sensors,”
IEEE J. Solid-State Circuits, vol. 35, pp. 724–731, May 2000.
[5] [Online]. Available: http://www.soton.ac.uk/~rmc1/robotics/artactile.htm
697
[6] B. V. Jayawant, M. A. Onorio, and J. D. MacWatson, “Robot tactile
sensing: A new array sensor,” in Robot Sensors II Tactile and Non-Vision, A. Pugh, Ed. London, U.K.: IFS, 1986, pp. 199–205.
[7] K. Tanie, “Advances in Tactile sensors for robotics,” in Proc. 6th Sensor
Symp., 1986, pp. 63–68.
[8] L. Liu, X. Zheng, and Z. Li, “An array tactile sensor with piezoresistive single crystal silicon diaphragm,” Sens. Actuators A, vol. 32, pp.
193–196, 1993.
[9] G. Murali Krishna, “Development of tactile sensors based on piezoelectric resonance and surface acoustic wave techniques,” M.Sc. (Engg.)
thesis, Indian Inst. Science, Bangalore, India, 2001.
[10] M. R. Neumann, Medical Progress Through Technology. New York:
Springer, 1982, vol. 9.
[11] S. Omata and Y. Terunuma, “New tactile sensor like human hand and its
applications,” Sens. Actuators A, vol. 35, pp. 9–15, 1992.
[12] P. Pagh, W. B. Heginbotham, and C. J. Page, “Novel techniques for tactile sensing in a three dimensional environment,” Ind. Robot., pp. 35–48,
Mar. 1977.
[13] P. Li and Y. Wen, “An arbitrarily distributed tactile piezoelectric sensor
array,” Sens. Actuators A, vol. 65, pp. 141–146, 1998.
[14] J. Rebman and K. A. Morris, “A tactile sensor with electro optic transduction,” in Proc. 3rd Int. Conf. Robot Vision and Sensory Controls,
Cambridge, MA, Nov. 1982, pp. 210–216.
[15] R. Adler and P. J. Desmares, “An economical touch panel using SAW
absorption,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol.
UFFC-34, pp. 195–201, Jan. 1987.
[16] J. Sorab, R. H. Allen, and B. Gonik, “Tactile sensory monitoring of clinician-applied forces during delivery of new borns,” IEEE Trans. Biomed.
Eng., vol. 35, pp. 1090–1093, Dec. 1998.
[17] S. Sugiyama, K. Kawahata, M. Yonendu, and I. Igarashi, “Tactile image
detection using 1 K-element silicon pressure sensor array,” Sens. Actuators A, vol. 21–23, pp. 397–400, 1990.
G. Murali Krishna was born in 1975 in India. He
received the M.Sc. degree in electronics from the
School of Physics, University of Hyderabad, India,
and the M.Sc. (Engg.) degree from the Department
of Instrumentation, Indian Institute of Science, Bangalore, where he was involved in the development of
tactile sensors. His thesis was titled “Development of
tactile sensors based on piezoelectric resonance and
surface acoustic waves.” He is currently pursuing
the Ph.D.degree in nanotechnology-based cantilever
sensors at the Institute of Physics, University of
Basel, Basel, Switzerland.
K. Rajanna was born in Karnataka, India. He
received the M.Sc. degree in physics from the
University of Mysore, India, in 1976 and the
M.Sc.(Engg.) and Ph.D. degrees from the Indian
Institute of Science (IISc), Bangalore, in 1988 and
1993, respectively.
At present, he is an Associate Professor at the
Department of Instrumentation, IISc. He was a Visiting Fellow of the Japan Society for the Promotion
of Science (JSPS) twice from 1997 to 1998 at the
Toyohashi University of Technology, Japan. He was
also a Visiting Professor at Tohoku University, Japan, during 2000 and 2002.
Recently, he was invited to serve on the program committee of the Asia-Pacific
Conference of Transducers and Micro-Nano Technology 2004 (APCOT’04) to
be held in Sapporo, Japan. His current areas of interest include thin-film sensors
and microsensors (including MEMS), vacuum and thin-film technology, and
novel materials engineering for sensor applications.
Dr. Rajanna served as a member of the Editorial Review Committee of the
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT.